Research Papers

Fall on Backpack: Damage Minimization of Humanoid Robots by Falling on Targeted Body Segments

[+] Author and Article Information
Sung-Hee Lee

School of Information and Communications,
Gwangju Institute of Science and Technology,
Gwangju, South Korea 500-712
e-mail: shl@gist.ac.kr

Ambarish Goswami

Honda Research Institute USA,
Mountain View, CA 94043
e-mail: agoswami@honda-ri.com

Our fall controller tries to make a point contact rather than an edge contact between the foot and the ground. The assumption on zero rotational friction simplifies the dynamic analysis of the falling robot, but the effect of the friction might not be negligible in reality.

[·] denotes a 3×3 skew symmetric matrix representing a vector cross product as in [a]b=a×b.

1Corresponding author.

Contributed by the Design Engineering Division of ASME for publication in the JOURNAL OF COMPUTATIONAL AND NONLINEAR DYNAMICS. Manuscript received August 30, 2011; final manuscript received April 18, 2012; published online July 23, 2012. Assoc. Editor: Arend L. Schwab.

J. Comput. Nonlinear Dynam 8(2), 021005 (Jul 23, 2012) (10 pages) Paper No: CND-11-1143; doi: 10.1115/1.4006783 History: Received August 30, 2011; Revised April 18, 2012

Safety and robustness will become critical issues when humanoid robots start sharing human environments in the future. In physically interactive human environments, a catastrophic fall is a major threat to the safety and smooth operation of humanoid robots. It is, therefore, imperative that humanoid robots be equipped with a comprehensive fall management strategy. This paper deals with the problem of reducing the impact damage to a robot associated with a fall. A common approach is to employ damage-resistant design and apply impact-absorbing material to robot limbs, such as the backpack and knee, that are particularly prone to fall related impacts. In this paper, we select the backpack to be the most preferred body segment to experience an impact. We proceed to propose a control strategy that attempts to reorient the robot during the fall such that it impacts the ground with its backpack. We show that the robot can fall on the backpack even when it starts falling sideways. This is achieved by generating and redistributing angular momentum among the robot limbs through dynamic coupling. The planning and control algorithms for a fall are demonstrated in simulation.

Copyright © 2013 by ASME
Your Session has timed out. Please sign back in to continue.



Grahic Jump Location
Fig. 1

Backpack is rigidly attached to the back of the trunk (left), and we aim to control a falling robot to touch down with a specific body part, in our case the backpack (right)

Grahic Jump Location
Fig. 2

The strategies for falling on the backpack are illustrated. (a) By counter-rotating the swing leg, the robot controls the spin of the trunk. (b) Additionally, the robot bends the trunk backward to increase the possibility of touching down with the backpack. The points G and P denote CoM and CoP, respectively.

Grahic Jump Location
Fig. 3

The Spin frame is located at the CoM and is defined by the axes {t,b,s}. The spin axis s coincides with the lean line PG→. The frontal axis t is the intersection of the sagittal plane of the trunk, and the plane normal to s (translucent circle). The bending direction b is given by s×t.

Grahic Jump Location
Fig. 4

Various configurations of the backpack when a robot successfully touches down with the backpack. Flat contact (middle) may reduce impact at collision, but it may not be achieved in practice because the backpack may have a curved contour. We do not care which point of the backpack collides with the ground.

Grahic Jump Location
Fig. 5

A falling robot is modeled as a single rigid body by locking all the joints and is simulated until its CoM drops down to a certain height ht, then the time to fall tf and the vector n, the normal direction pointing backward from the backpack, are measured. Note that n is not necessarily orthogonal to the lean line GP (see Fig. 6 for n). The direction of n' is the desired value of n determined by rotating n about the lean line such that the plane made by GP and n' is orthogonal to the ground plane.

Grahic Jump Location
Fig. 6

The figure illustrates how the bending strategy helps falling on the backpack. The boxes represent the configurations of the backpack at different instances during the fall: A: when a robot is standing upright. B: when the fall controller is activated. C: when the robot hits the ground while only the spinning strategy is employed. D: when the robot hits the ground while both the spinning and bending strategies are employed. The vector n represents the normal direction of the backpack, pointing backward. Including the bending strategy makes n more orthogonal to the ground plane.

Grahic Jump Location
Fig. 7

An external force of 160 N applied at the CoM of the robot to its left for 300 ms makes the robot fall. Top row: The robot locks all the joints without triggering the fall controller. It falls sideways. Bottom row: The robot engages only the spinning strategy. It lifts and rotate its right leg to control the spin angle of the trunk. The trunk bending strategy is not used. The trunk spins noticeably, but not enough to fall on the backpack. As a result, the robot's hand hits the ground first.

Grahic Jump Location
Fig. 8

Under the same external force as in Fig. 7, the robot uses both the spinning and bending strategies. It successfully touches the ground with the backpack. The top, middle, and bottom rows show the side, top, and front views, respectively, of the simulation.

Grahic Jump Location
Fig. 9

Plots of CoM (dotted line marked with circles), frontal axis t of the spin frame (gray arrows), the position of the trunk frame (solid line marked with squares), and the normal vector n to the backpack (black arrows, see Fig. 5 for n) for the simulation of Fig. 8. Frontal axes and the normal vectors are drawn at every 50 ms. The inset robot figure in each graph indicates the viewpoint in which the graph is drawn. The vector n is not orienting completely downward at touchdown time, but the robot still manages to touch the ground with the backpack.

Grahic Jump Location
Fig. 10

Spin velocity (top) and the angle between the normal direction of the backpack and the downward normal direction of the ground (bottom) of the fall control experiment in Fig. 8

Grahic Jump Location
Fig. 11

Angular momentum of the robot in the experiment of Fig. 8 Top: angular momentum with respect to the spin frame. Bottom: Decomposition of the spin component of the angular momentum (dotted line, middle) into two sources, one due to swing leg (solid line, top) and the other due to all other body parts (dotted line, bottom).

Grahic Jump Location
Fig. 12

Estimation errors, calculated as the difference between the actual and the estimated touchdown times, in each of the planning stages of the experiment in Fig. 8

Grahic Jump Location
Fig. 13

Experiments with different push directions




Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In